Diverting the Diagnosis: A Case Report of Hemodialysis Masking the Etiology of Hyperammonemia

Abstract

Rationale:

Hyperammonemia in patients receiving hemodialysis is uncommon but poses a significant clinical challenge due to the effective clearance of ammonia by dialysis, which can obscure the underlying cause. Recognizing atypical etiologies is crucial for appropriate management.

Presenting Concerns of the Patient:

A 59-year-old man being treated with hemodialysis presented with altered level of consciousness and recurrent hyperammonemia. Despite previous episodes of hyperammonemia, the etiology of his intermittently elevated ammonia remained unclear and was initially attributed to his kidney failure.

Diagnoses:

Initial assessments, including liver function tests, abdominal ultrasound, medication review, and genetic screening for urea cycle disorders, were unremarkable. Upon recurrence of symptoms with hyperammonemia, a computed tomography scan was performed which revealed a large portosystemic shunt between the splenic vein and right common iliac vein.

Interventions:

The patient underwent embolization of the identified portosystemic shunt.

Outcomes:

Following embolization of the shunt, the patient’s hyperammonemia and encephalopathy resolved, with no further recurrences.

Novel Findings:

This case illustrates the challenges of determining the etiology of hyperammonemia in patients treated with hemodialysis due to the dialysis clearance of ammonia. Portosystemic shunts cause hyperammonemia by bypassing the liver’s ammonia-detoxification pathways, and their effects may be paradoxically exacerbated immediately after dialysis due to dialysis-related hemodynamic changes. We emphasize the importance of investigating hyperammonemia as a cause of altered level of consciousness among patients being treated with hemodialysis and considering anatomical shunting in the differential diagnosis.

Introduction

Ammonia is produced in the body primarily as a byproduct of amino acid metabolism. Several organ systems are involved in its production and removal, including the liver where ammonia is detoxified via the urea cycle, the kidneys through glutamine metabolism and urea excretion, and the intestines where bacterial urease enzymes convert urea to ammonia. ¹ Although the reduced capacity for renal clearance of ammonia (compared to hepatic metabolism of ammonia) may predispose patients with kidney failure to hyperammonemia, this is not commonly observed as dialysis effectively clears ammonia.^2,3

Presenting Concerns

A 59-year-old man presented to hospital for assessment of altered level of consciousness following routine outpatient dialysis. He had been receiving intermittent hemodialysis for 4 months for treatment of kidney failure secondary to diabetes. His medical profile also included a left thalamic hemorrhagic stroke 6 years prior and seizure disorder maintained on levetiracetam and phenytoin.

He had a similar presentation for altered level of consciousness 4 months prior where he was found to have hyperammonemia without a clear etiology. It improved with initiation of dialysis, lactulose, and rifaximin. His initial workup revealed normal liver function, unremarkable liver ultrasound, and a negative genetic panel for urea cycle disorders. He also developed seizures during this admission and was initiated on antiepileptic therapy. His hyperammonemia was attributed to his kidney failure, which was felt to be lowering his seizure threshold. He was continued on intermittent hemodialysis thrice weekly upon discharge.

Clinical Findings

During the current presentation, a physical exam revealed blood pressure of 159/97 mm Hg, heart rate of 102 beats per minute, respiratory rate at 26 breaths per minute, and oxygen saturation of 100% on room air. The patient was afebrile with a Glasgow Coma Scale score of 6 (E1V1M4). His neurological exam was notable for right facial droop and right arm fixed flexion (his known baseline), bilateral horizontal nystagmus, and positive right plantar reflex. His exam was negative for asterixis, tremors, and clonus, and his cardiovascular, respiratory, abdominal, and skin examinations were unremarkable.

Computed tomography (CT) angiogram was negative for an acute ischemic stroke or hemorrhage. He was treated with a loading dose of levetiracetam due to the concern for non-convulsive status epilepticus, and an urgent electroencephalogram (EEG) was arranged.

Timeline

Table 1 outlines the relevant clinical events and diagnostic workup leading up to and during the principal admission.

Table 1.

Timeline of Relevant Clinical Events, Starting Prior to Index Admission and Continuing Through to Discharge.

Date	Clinical event
Day −141	Admitted to hospital for altered level of consciousness with progressive kidney failure. Found to have hyperammonemia with normal liver function. Initiated on intermittent hemodialysis.
Day −138	Episode of acutely reduced level of consciousness with generalized tonic-clonic movements. Initiated on levetiracetam and phenytoin by neurology.
Day −117	Genetic testing for hyperammonemia and urea cycle disorders did not detect any known disease-causing or rare variants.
Day −115	Discharged from hospital to long-term care facility on intermittent hemodialysis three times weekly.
Day 0	Presented to hospital for altered level of consciousness during outpatient hemodialysis. Investigations demonstrated hyperammonemia and unchanged CT-A head and neck.
	Initial laboratory values
	Hgb (g/L): 122 (135-175)	HCO3 (mmol/L): 24 (20-32)	Ammonia (μmol/L): 145 (20-50)
	WBC (109/L): 4.4 (4-11)	Glucose (mmol/L): 8.1	ALT (U/L): 203 (<70)
	Plt (109/L): 256 (140-400)	Ca (mmol/L): 2.17 (2.1-2.6)	AST (U/L): 82 (<55)
	INR: 0.9 (0.8-1.2)	Phos (mmol/L): 1.11 (0.7-1.5)	ALP (U/L): 451 (40-120)
	Na (mmol/L): 136 (135-145)	Mg (mmol/L): 1.14 (0.7-1.0)	GGT (U/L): 1183 (<80)
	K (mmol/L): 3.9 (3.5-5.0)	Creatinine (μmol/L): 305	Bilirubin (μmol/L): 7 (<20)
	Cl (mmol/L): 99 (98-112)	Urea (mmol/L): 14.9 (3-9)	Albumin (g/L): 28 (30-45)
Day 3	Large portosystemic shunt identified on CT abdomen pelvis.
Day 6	IR embolization of portosystemic shunt.
Day 8	Returned to cognitive and functional baseline. Discharged back to long-term care facility.

Note. CT = computed tomography; CT-A = computed tomography angiography; IR = interventional radiology. Hgb = hemoglobin; WBC = white blood cell count; Plt = platelet; INR = international normalized ratio; Na = sodium; K = potassium; Cl = chloride; HCO3 = bicarbonate; Ca = calcium; Phos = phosphate; Mg = magnesium; ALT = alanine aminotransferase; AST = aspartate aminotransferase; ALP = alkaline phosphatase; GGT = gamma-glutamyl transferase.

Diagnostic Focus and Assessment

Relevant Investigations

Laboratory results

Initial laboratory tests were pertinent for elevated ammonia at 145 μmol/L (reference range 20-50), and mixed liver enzyme elevation, which are summarized in Table 1.

Imaging and diagnostics

Day 0: CT/CT-A head and neck—no acute findings with patent intracranial vessels, redemonstration of left thalamic infarct.

Day 0: EEG—moderate to severe persistent generalized slowing indicating underlying global cerebral dysfunction. No definite epileptiform discharges or seizures.

Day 1: Abdominal ultrasound with doppler—morphologically normal appearing liver. Patent portal and hepatic veins with normal flow.

Day 3: CT abdomen pelvis enhanced—large portosystemic shunt coursing between splenic vein and right common iliac vein, up to 1.6cm in diameter.

Diagnostic Reasoning

After discontinuing phenytoin and initiating clobazam, his liver enzymes normalized. Genetic testing for urea cycle disorders, urine and blood cultures, and a medication review did not identify the cause of hyperammonemia. Due to the resolution of the elevated liver enzymes, occurrence of hyperammonemia before the transient rise in liver enzymes, and absence of cirrhotic features on abdominal ultrasound, liver disease was felt unlikely to be the cause of the hyperammonemia. Next, attention was turned to investigating a possible portosystemic shunt.

Therapeutic Focus and Assessment

The patient improved after additional hemodialysis, increased lactulose, and antiepileptic loading. His EEG showed no seizures, and no other cause for his altered level of consciousness was found. Given the persistence of hyperammonemia despite dialysis and unremarkable liver disease and genetic testing, a CT scan was ordered to assess for an anatomical shunt bypassing the hepatic circulation. This revealed a large portosystemic shunt between the splenic vein and the right common iliac vein, which was successfully treated with IR embolization and resulted in sustained normalization of his ammonia levels.

Follow-up and Outcomes

The patient returned to his baseline cognition and was discharged to his long-term care facility on thrice weekly hemodialysis. He has remained out of hospital since. There are ongoing discussions as to whether his antiepileptic medications should be tapered since the hyperammonemia, which may have lowered his seizure threshold, has resolved.

Discussion

We report a case of hyperammonemia due to a large portosystemic shunt in a patient treated with hemodialysis. This case highlights the potential delays in diagnosis that patients treated with hemodialysis may experience, particularly when clinical abnormalities are initially attributed to their underlying kidney failure.

Treatment with hemodialysis likely delayed the diagnosis in this case, as dialysis effectively removes ammonia from the bloodstream. Ammonia’s low molecular weight (17 mg/mmol), water solubility, and minimal protein binding enhance its clearance by hemodialysis.^3,4 Ammonia clearance increases with higher blood flow rate, dialysate flow rate, and dialyzer surface area; when these are optimized, dialysis clearance can approach the liver’s normal ammonia clearance capacity.^3,4 Consequently, an underlying disorder causing hyperammonemia and associated encephalopathy may be less apparent in patients treated with dialysis, as ammonia levels and clinical presentation fluctuates with and between dialysis treatments.

The differential diagnosis for hyperammonemia can be categorized mechanistically: hepatic versus non-hepatic causes, genetic versus non-genetic factors, and increased production versus impaired clearance, as summarized in Figure 1.^5,6 Anatomical portosystemic shunts, impair ammonia clearance, and can be classified as either hepatic or non-hepatic. Although these may occur in the absence of liver disease (such as in our case), this is less common. ⁶

Figure 1.

Differential diagnosis of hyperammonemia, organized by mechanisms of increased ammonia production and decreased ammonia clearance.^5,6

To our knowledge, only three other case reports describe hyperammonemia due to portosystemic shunts in patients treated with dialysis. One report describes a patient with an acute altered level of consciousness after hemodialysis, later found to have a large portosystemic shunt between the left gastric and left renal veins, which was treated with surgical ligation. ⁷ Another case involved a patient on peritoneal dialysis who developed encephalopathy with elevated ammonia several months after starting dialysis and was found to have a shunt from the left gastric vein to the azygous vein. The patient’s symptoms partially improved with transition to hemodialysis, but resolved completely following surgical ligation of the shunt. ⁸ A third report described two patients with portosystemic shunts: one had a left gastric to left renal vein shunt with improved encephalopathy after dialysis initiation, while the other had a splenic to left renal vein shunt but experienced recurrent encephalopathy episodes after dialysis and was subsequently treated with balloon-occluded retrograde transvenous obliteration. ⁹

The paradoxical increase in ammonia and worsening symptoms after large volume removal during dialysis, as seen in our case and others, has been reported in the literature. This is hypothesized to result from increased portal venous shunt flow following dialysis due to a significant decrease in intravascular pressure with high ultrafiltration. ⁷ Another study, examining patients without portosystemic shunts, also found a subset of patients experienced increased post-dialysis ammonia levels, which correlated with greater ultrafiltration and lower mean arterial pressure, suggesting that reduced hepatic perfusion may transiently impair ammonia clearance. ¹⁰ In our case, the patient initially improved with dialysis but had another acute encephalopathy episode after hemodialysis. His initial improvement may also have been attributable to administration of lactulose and rifaximin, which reduce ammonia by promoting intestinal excretion and decreasing intestinal production, respectively. As with previous cases, the definitive therapy for our patient was closure of the shunt.

In summary, we emphasize the importance of considering hyperammonemia in the workup of acute encephalopathy in patients treated with dialysis, even with normal liver function. There are various mechanisms for hyperammonemia in the absence of liver disease, such as portosystemic shunting as in our case. Kidney failure presents a unique challenge in diagnosing hyperammonemia associated encephalopathy as hemodialysis effectively clears ammonia from the bloodstream. Furthermore, patients with portosystemic shunts may clinically worsen after dialysis treatments despite clearance of ammonia, due to increased portal venous shunting from dialysis-associated hemodynamic changes. Our case highlights the unique clinical presentation of hyperammonemia among patients treated with dialysis and emphasizes the importance of dedicated cross-sectional imaging to detect portosystemic shunts when evaluating hyperammonemia. Recognizing these pathophysiological mechanisms is crucial for timely diagnosis and appropriate management of these patients.